01/AI Agent · Commerce
2026ReplyFlows
A WhatsApp agent connected to a live Shopify store: it answers product questions, checks stock, looks up orders and keeps selling after hours.
- TypeScript
- React
- Node.js
- PostgreSQL

Mahmoud Atieh
AI Automation Engineer
Agents, retrieval systems and automated workflows — built as products that run in production, grounded in real data and connected to real tools.

Full-stack AI products I built end to end and deployed live to production.
Four core systems architected for production and verified in real use.
Production agents connected to real APIs, tools and databases — they resolve intent, call the right service dynamically, and answer from live data instead of scripted flows.
Grounded retrieval pipelines built for precision and zero hallucination: document OCR ingestion, dense vector & BM25 hybrid search, and cross-encoder re-ranking.
Resilient API-driven automations with webhooks, OAuth and scheduled jobs — taking complex manual processes off people's desks with battle-tested error handling.
The robust engineering that makes AI features dependable: type-safe backends, relational databases, custom dashboards, containerized deployment, and live monitoring.
Building retrieval systems, agents and automated workflows, with 1+ year of experience in AI automation.
I build the product around them too — the API, the database, the dashboard and the deployment. A model that performs in a notebook is a long way from a system someone can depend on at nine in the morning.
I care about grounding. In every project here, an answer traces back to something real — a page in a document, a product record, an order. That constraint shapes the architecture more than the model choice does.
Send the process as it actually happens — the tools, the volume, what goes wrong. I'll tell you whether it's worth automating and what I'd build first.